# Manipulation et analyse des données
import numpy as np
import pandas as pd

# Visualisation
import matplotlib.pyplot as plt
import seaborn as sns
import matplotlib.patches as mpatches
import plotly.figure_factory as ff
import plotly.graph_objs as go
from plotly.offline import iplot, init_notebook_mode
from plotly.subplots import make_subplots

init_notebook_mode(connected=True)


# Réduction de dimension
from sklearn.decomposition import PCA, TruncatedSVD
from sklearn.manifold import TSNE


# Modèles de classification
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier

# Préparation des données et évaluation
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split, KFold, StratifiedKFold
from sklearn.pipeline import make_pipeline
from sklearn.metrics import (
    accuracy_score,
    precision_score,
    recall_score,
    f1_score,
    roc_auc_score,
    classification_report
)

# Gestion du déséquilibre des classes
from imblearn.under_sampling import TomekLinks
from imblearn.over_sampling import SMOTE
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report, roc_auc_score
from sklearn.metrics import confusion_matrix


# Autres utilitaires
from collections import Counter
import time
import warnings
warnings.filterwarnings("ignore")

# Configuration des graphiques
plt.style.use("default")
sns.set_context("notebook")

# Lecture du fichier CSV
df = pd.read_csv("creditcard.csv")

# Affichage des premières lignes du jeu de données
df.head(10)

# Dimensions du jeu de données

print(f"Dimensions du jeu de données : {df.shape[0]} lignes et {df.shape[1]} colonnes")

Dimensions du jeu de données : 284807 lignes et 31 colonnes

# Liste des colonnes du jeu de données
df.columns

Index(['Time', 'V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10',
       'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20',
       'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount',
       'Class'],
      dtype='object')

# Informations générales sur la structure et les types de données
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 284807 entries, 0 to 284806
Data columns (total 31 columns):
 #   Column  Non-Null Count   Dtype  
---  ------  --------------   -----  
 0   Time    284807 non-null  float64
 1   V1      284807 non-null  float64
 2   V2      284807 non-null  float64
 3   V3      284807 non-null  float64
 4   V4      284807 non-null  float64
 5   V5      284807 non-null  float64
 6   V6      284807 non-null  float64
 7   V7      284807 non-null  float64
 8   V8      284807 non-null  float64
 9   V9      284807 non-null  float64
 10  V10     284807 non-null  float64
 11  V11     284807 non-null  float64
 12  V12     284807 non-null  float64
 13  V13     284807 non-null  float64
 14  V14     284807 non-null  float64
 15  V15     284807 non-null  float64
 16  V16     284807 non-null  float64
 17  V17     284807 non-null  float64
 18  V18     284807 non-null  float64
 19  V19     284807 non-null  float64
 20  V20     284807 non-null  float64
 21  V21     284807 non-null  float64
 22  V22     284807 non-null  float64
 23  V23     284807 non-null  float64
 24  V24     284807 non-null  float64
 25  V25     284807 non-null  float64
 26  V26     284807 non-null  float64
 27  V27     284807 non-null  float64
 28  V28     284807 non-null  float64
 29  Amount  284807 non-null  float64
 30  Class   284807 non-null  int64  
dtypes: float64(30), int64(1)
memory usage: 67.4 MB

# Statistiques descriptives des variables numériques
df.describe()

### Analyse des valeurs manquantes

total = df.isnull().sum().sort_values(ascending = False)
percent = (df.isnull().sum()/df.isnull().count()*100).sort_values(ascending = False)
pd.concat([total, percent], axis=1, keys=['Total', 'Pourcentage']).transpose()

fig, ax = plt.subplots(1, 2, figsize=(18,4))

amount_val = df['Amount'].values
time_val = df['Time'].values

sns.distplot(amount_val, ax=ax[0], color='r')
ax[0].set_title('Distribution de la variable Amount', fontsize=14)
ax[0].set_xlim([min(amount_val), max(amount_val)])

sns.distplot(time_val, ax=ax[1], color='b')
ax[1].set_title('Distribution de la variable Time', fontsize=14)
ax[1].set_xlim([min(time_val), max(time_val)])

(0.0, 172792.0)

# Calcul de l'asymétrie pour les variables Time et Amount

skew_df = df[['Time', 'Amount']].skew().to_frame(name="Skewness")

skew_df["Interprétation"] = [
    "≈ 0 (distribution quasi symétrique)",
    ">> 0 (forte asymétrie positive)"
]

skew_df

# Boxplots pour détecter les valeurs aberrantes

# Palette de couleurs personnalisée
colors = ["#8DA3F1FF", "#CD2121"]  # bleu, rouge

fig, ax = plt.subplots(1, 2, figsize=(14, 4))

sns.boxplot(x=df['Amount'], ax=ax[0], color=colors[1])
ax[0].set_title("Boxplot de la variable Amount")

sns.boxplot(x=df['Time'], ax=ax[1], color=colors[0])
ax[1].set_title("Boxplot de la variable Time")

plt.tight_layout()
plt.show()

# Comptage des classes
class_counts = df['Class'].value_counts()

# Pourcentage des classes
class_percent = (df['Class'].value_counts(normalize=True) * 100).round(2)

# Création du DataFrame
class_df = pd.concat([class_counts, class_percent], axis=1)
class_df.columns = ['Nombre', 'Pourcentage (%)']

# Renommer l'index
class_df.index = ['Class 0', 'Class 1']

# Ajout de l'interprétation
class_df['Interprétation'] = [
    "Classe majoritaire (transactions légitimes)",
    "Classe minoritaire (transactions frauduleuses)"
]

class_df

# Comptage des classes
temp = df["Class"].value_counts()

# Création d'un DataFrame pour Plotly
class_df = pd.DataFrame({
    'Class': temp.index,
    'Valeurs': temp.values
})

# Création du graphique en barres
trace = go.Bar(
    x=class_df['Class'],
    y=class_df['Valeurs'],
    name="Déséquilibre des classes (0 = légitime, 1 = fraude)",
    marker=dict(color=["#DF0101", "#3101DF"]),
    text=class_df['Valeurs'],
    textposition='auto'
)

data = [trace]

layout = go.Layout(
    title="Déséquilibre des classes : transactions légitimes vs frauduleuses",
    xaxis=dict(title="Classe (0 = légitime, 1 = fraude)"),
    yaxis=dict(title="Nombre de transactions"),
    hovermode='closest',
    width=600
)

fig = go.Figure(data=data, layout=layout)

iplot(fig)

# Séparation des transactions selon la classe
class_0 = df.loc[df['Class'] == 0, 'Time']
class_1 = df.loc[df['Class'] == 1, 'Time']

# Données pour le graphique
hist_data = [class_0, class_1]
group_labels = ['Class 0 (Légitime)', 'Class 1 (Fraude)']

# Création du graphique de densité
fig = ff.create_distplot(
    hist_data,
    group_labels,
    show_hist=False,
    show_rug=False
)

fig.update_layout(
    title='Densité temporelle des transactions (Time)',
    xaxis=dict(title='Temps écoulé (secondes)')
)

iplot(fig)

# Transformation logarithmique de la variable Amount
df['Amount_log'] = np.log1p(df['Amount'])

# Vérifier la présence des variables transformées
df[['Amount', 'Amount_log']].head(10)

# Asymétrie avant et après transformation

# Calcul de l'asymétrie
skew_amount = df[['Amount', 'Amount_log']].skew()

# Création du DataFrame récapitulatif
skew_df = skew_amount.to_frame(name="Skewness")

# Ajout de l'interprétation
skew_df['Interprétation'] = [
    ">> 0 : forte asymétrie positive",
    "≈ 0 : distribution quasi symétrique"
]

skew_df

marker_color1 = "rgba(214, 4, 4, 0.6)"   # rouge semi-transparent
marker_color2 = "rgba(1, 1, 223, 0.6)"   # bleu semi-transparent

fig = make_subplots(
    rows=1, cols=2,
    subplot_titles=(
        "Distribution de la variable Amount",
        "Distribution de la variable Amount_log"
    )
)

# Histogramme Amount (avant)
fig.add_trace(
    go.Histogram(
        x=df['Amount'],
        histnorm='probability density',
        nbinsx=50,
        marker_color=marker_color1,
        opacity=0.6,
        name='Amount'
    ),
    row=1, col=1
)

# Histogramme Amount_log (après)
fig.add_trace(
    go.Histogram(
        x=df['Amount_log'],
        histnorm='probability density',
        nbinsx=50,
        marker_color=marker_color2,
        opacity=0.6,
        name='Amount_log'
    ),
    row=1, col=2
)

# Mise en page
fig.update_layout(
    height=400,
    width=900,
    showlegend=False,
    bargap=0.05,
    xaxis_title="Amount",
    xaxis2_title="Amount_log",
    yaxis_title="Density",
    title_text="Comparaison des distributions avant et après transformation logarithmique"
)

iplot(fig)

# Transformation de la variable Time en heures
df['Time_hours'] = df['Time'] / 3600

# Vérifier la présence des variables transformées
df[['Time', 'Time_hours']].head(10)

# Vérification de la transformation temporelle
df[['Time', 'Time_hours']].describe()

# Standardisation du montant transformé
scaler = StandardScaler()
df['Amount_scaled'] = scaler.fit_transform(df[['Amount_log']])

# Création d'une variable représentant la période de la journée
df['Time_period'] = pd.cut(
    df['Time_hours'] % 24,
    bins=[0, 6, 12, 18, 24],
    labels=['Nuit', 'Matin', 'Après-midi', 'Soir'],
    include_lowest=True
)

df[['Amount', 'Amount_log', 'Amount_scaled', 'Time_hours', 'Time_period']].head()

# Encodage des variables catégorielles (one-hot encoding)

df_encoded = pd.get_dummies(df, columns=['Time_period'], drop_first=True)

df_encoded[['Amount', 'Amount_log', 'Amount_scaled', 'Time_hours', 'Time_period_Matin', 'Time_period_Après-midi', 'Time_period_Soir']].head()

# supprimer les variables d'origine
df_encoded.drop(['Amount', 'Amount_log', 'Time','Time_hours'], axis=1, inplace=True) 

df_encoded.head()

# Préparation des données pour la modélisation
# Séparation des caractéristiques (X) et de la cible (y)
X = df_encoded.drop(columns=['Class'])
y = df_encoded['Class']

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.3, random_state=42, stratify=y
)

# Gestion du déséquilibre des classes avec Tomek Links

tomek = TomekLinks()
X_train_tl, y_train_tl = tomek.fit_resample(X_train, y_train)

model_tl = LogisticRegression(max_iter=1000)
model_tl.fit(X_train_tl, y_train_tl)

y_pred_tl = model_tl.predict(X_test)
y_prob_tl = model_tl.predict_proba(X_test)[:, 1]

print("Résultats avec Tomek Links")
print(classification_report(y_test, y_pred_tl))
print("ROC-AUC :", roc_auc_score(y_test, y_prob_tl))

Résultats avec Tomek Links
              precision    recall  f1-score   support

           0       1.00      1.00      1.00     85295
           1       0.85      0.65      0.74       148

    accuracy                           1.00     85443
   macro avg       0.92      0.82      0.87     85443
weighted avg       1.00      1.00      1.00     85443

ROC-AUC : 0.953274169297969

# Calcul de la matrice de confusion
cm = confusion_matrix(y_test, y_pred_tl)

# Création du heatmap Plotly
fig = go.Figure(
    data=go.Heatmap(
        z=cm,
        x=['Prédit : Légitime (0)', 'Prédit : Fraude (1)'],
        y=['Réel : Légitime (0)', 'Réel : Fraude (1)'],
        colorscale='Blues',
        showscale=True,
        text=cm,
        texttemplate="%{text}",
        hoverinfo="text"
    )
)

# Mise en forme
fig.update_layout(
    title="Matrice de confusion – Tomek Links",
    xaxis_title="Classe prédite",
    yaxis_title="Classe réelle",
    width=500,
    height=400
)

iplot(fig)

# Gestion du déséquilibre des classes avec SMOTE
smote = SMOTE(random_state=42)
X_train_smote, y_train_smote = smote.fit_resample(X_train, y_train)

model_smote = LogisticRegression(max_iter=1000)
model_smote.fit(X_train_smote, y_train_smote)

y_pred_smote = model_smote.predict(X_test)
y_prob_smote = model_smote.predict_proba(X_test)[:, 1]

print("Résultats avec SMOTE")
print(classification_report(y_test, y_pred_smote))
print("ROC-AUC :", roc_auc_score(y_test, y_prob_smote))

Résultats avec SMOTE
              precision    recall  f1-score   support

           0       1.00      0.99      0.99     85295
           1       0.09      0.86      0.17       148

    accuracy                           0.99     85443
   macro avg       0.55      0.93      0.58     85443
weighted avg       1.00      0.99      0.99     85443

ROC-AUC : 0.9658917065256827

# Calcul de la matrice de confusion pour SMOTE
cm_sm = confusion_matrix(y_test, y_pred_smote)

fig = go.Figure(
    data=go.Heatmap(
        z=cm_sm,
        x=['Prédit : Légitime (0)', 'Prédit : Fraude (1)'],
        y=['Réel : Légitime (0)', 'Réel : Fraude (1)'],
        colorscale='Oranges',
        showscale=True,
        text=cm_sm,
        texttemplate="%{text}",
        hoverinfo="text"
    )
)

fig.update_layout(
    title="Matrice de confusion – SMOTE",
    xaxis_title="Classe prédite",
    yaxis_title="Classe réelle",
    width=500,
    height=400
)

iplot(fig)

# --- Tomek Links ---
precision_tl = precision_score(y_test, y_pred_tl, pos_label=1)
recall_tl    = recall_score(y_test, y_pred_tl, pos_label=1)
f1_tl        = f1_score(y_test, y_pred_tl, pos_label=1)
roc_tl       = roc_auc_score(y_test, y_prob_tl)

# --- SMOTE ---
precision_smote = precision_score(y_test, y_pred_smote, pos_label=1)
recall_smote    = recall_score(y_test, y_pred_smote, pos_label=1)
f1_smote        = f1_score(y_test, y_pred_smote, pos_label=1)
roc_smote       = roc_auc_score(y_test, y_prob_smote)

# DataFrame comparaison
compare_df = pd.DataFrame({
    "Méthode": ["Tomek Links", "SMOTE"],
    "Precision (Fraude=1)": [precision_tl, precision_smote],
    "Recall (Fraude=1)":    [recall_tl, recall_smote],
    "F1-score (Fraude=1)":  [f1_tl, f1_smote],
    "ROC-AUC":              [roc_tl, roc_smote]
})

compare_df.set_index("Méthode").round(3)

marker_color1 = "rgba(314, 4, 4, 0.6)"   # rouge semi-transparent
marker_color2 = "rgba(1, 1, 223, 0.6)"   # bleu semi-transparent

metrics = ["Precision (Fraude=1)", "Recall (Fraude=1)", "F1-score (Fraude=1)", "ROC-AUC"]

fig = go.Figure()

fig.add_trace(go.Bar(
    name="Tomek Links",
    x=metrics,
    y=[precision_tl, recall_tl, f1_tl, roc_tl],
    marker_color=marker_color1
))

fig.add_trace(go.Bar(
    name="SMOTE",
    x=metrics,
    y=[precision_smote, recall_smote, f1_smote, roc_smote],
    marker_color=marker_color2
))

fig.update_layout(
    title="Comparaison des performances : Tomek Links vs SMOTE",
    barmode="group",
    yaxis_title="Score",
    xaxis_title="Métriques",
    width=900,
    height=450
)

iplot(fig)

# Dimensions du jeu de données final
print("Dimensions du jeu de données final :", df_encoded.shape)

# Aperçu des premières lignes
df_encoded.head()

Dimensions du jeu de données final : (284807, 33)

# Vérification des types de données
df_encoded.dtypes

V1                        float64
V2                        float64
V3                        float64
V4                        float64
V5                        float64
V6                        float64
V7                        float64
V8                        float64
V9                        float64
V10                       float64
V11                       float64
V12                       float64
V13                       float64
V14                       float64
V15                       float64
V16                       float64
V17                       float64
V18                       float64
V19                       float64
V20                       float64
V21                       float64
V22                       float64
V23                       float64
V24                       float64
V25                       float64
V26                       float64
V27                       float64
V28                       float64
Class                       int64
Amount_scaled             float64
Time_period_Matin            bool
Time_period_Après-midi       bool
Time_period_Soir             bool
dtype: object

### Analyse des valeurs manquantes

total = df.isnull().sum().sort_values(ascending = False)
percent = (df.isnull().sum()/df.isnull().count()*100).sort_values(ascending = False)
pd.concat([total, percent], axis=1, keys=['Total', 'Pourcentage']).transpose()

df_encoded.describe()

skew_value = df_encoded[['Amount_scaled']].skew().iloc[0]
skew_value

0.16270292774164002

fig = go.Figure()

# Histogramme
fig.add_trace(go.Histogram(
    x=df_encoded['Amount_scaled'],
    nbinsx=50,
    marker_color="#3101DF",
    opacity=0.7,
    name="Amount_scaled",
    histnorm="probability density"
))

# Ligne verticale (moyenne)
fig.add_trace(go.Scatter(
    x=[df_encoded['Amount_scaled'].mean(), df_encoded['Amount_scaled'].mean()],
    y=[0, 0.5],
    mode="lines",
    line=dict(color="#DF0101", dash="dash"),
    name="Moyenne"
))

fig.update_layout(
    title=f"Distribution de Amount_scaled (Asymétrie = {skew_value:.3f})",
    xaxis_title="Amount_scaled",
    yaxis_title="Densité",
    width=700,
    height=450,
    showlegend=True
)

iplot(fig)

# Calcul de la matrice de corrélation
corr_matrix = df_encoded.corr()

# Création du heatmap interactif
fig = go.Figure(
    data=go.Heatmap(
        z=corr_matrix.values,
        x=corr_matrix.columns,
        y=corr_matrix.columns,
        colorscale='RdBu',
        zmin=-1,
        zmax=1,
        showscale=True,
        colorbar=dict(title="Corrélation")
    )
)

fig.update_layout(
    title="Matrice de corrélation des variables (df_encoded)",
    width=900,
    height=900,
    xaxis=dict(tickangle=45),
    yaxis=dict(autorange='reversed')
)

iplot(fig)

	Time	V1	V2	V3	V4	V5	V6	V7	V8	V9	...	V21	V22	V23	V24	V25	V26	V27	V28	Amount
0	0.0	-1.359807	-0.072781	2.536347	1.378155	-0.338321	0.462388	0.239599	0.098698	0.363787	...	-0.018307	0.277838	-0.110474	0.066928	0.128539	-0.189115	0.133558	-0.021053	149.62
1	0.0	1.191857	0.266151	0.166480	0.448154	0.060018	-0.082361	-0.078803	0.085102	-0.255425	...	-0.225775	-0.638672	0.101288	-0.339846	0.167170	0.125895	-0.008983	0.014724	2.69
2	1.0	-1.358354	-1.340163	1.773209	0.379780	-0.503198	1.800499	0.791461	0.247676	-1.514654	...	0.247998	0.771679	0.909412	-0.689281	-0.327642	-0.139097	-0.055353	-0.059752	378.66
3	1.0	-0.966272	-0.185226	1.792993	-0.863291	-0.010309	1.247203	0.237609	0.377436	-1.387024	...	-0.108300	0.005274	-0.190321	-1.175575	0.647376	-0.221929	0.062723	0.061458	123.50
4	2.0	-1.158233	0.877737	1.548718	0.403034	-0.407193	0.095921	0.592941	-0.270533	0.817739	...	-0.009431	0.798278	-0.137458	0.141267	-0.206010	0.502292	0.219422	0.215153	69.99
5	2.0	-0.425966	0.960523	1.141109	-0.168252	0.420987	-0.029728	0.476201	0.260314	-0.568671	...	-0.208254	-0.559825	-0.026398	-0.371427	-0.232794	0.105915	0.253844	0.081080	3.67
6	4.0	1.229658	0.141004	0.045371	1.202613	0.191881	0.272708	-0.005159	0.081213	0.464960	...	-0.167716	-0.270710	-0.154104	-0.780055	0.750137	-0.257237	0.034507	0.005168	4.99
7	7.0	-0.644269	1.417964	1.074380	-0.492199	0.948934	0.428118	1.120631	-3.807864	0.615375	...	1.943465	-1.015455	0.057504	-0.649709	-0.415267	-0.051634	-1.206921	-1.085339	40.80
8	7.0	-0.894286	0.286157	-0.113192	-0.271526	2.669599	3.721818	0.370145	0.851084	-0.392048	...	-0.073425	-0.268092	-0.204233	1.011592	0.373205	-0.384157	0.011747	0.142404	93.20
9	9.0	-0.338262	1.119593	1.044367	-0.222187	0.499361	-0.246761	0.651583	0.069539	-0.736727	...	-0.246914	-0.633753	-0.120794	-0.385050	-0.069733	0.094199	0.246219	0.083076	3.68

	Time	V1	V2	V3	V4	V5	V6	V7	V8	V9	...	V21	V22	V23	V24	V25	V26	V27	V28	Amount	Class
count	284807.000000	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	...	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	2.848070e+05	284807.000000	284807.000000
mean	94813.859575	1.168375e-15	3.416908e-16	-1.379537e-15	2.074095e-15	9.604066e-16	1.487313e-15	-5.556467e-16	1.213481e-16	-2.406331e-15	...	1.654067e-16	-3.568593e-16	2.578648e-16	4.473266e-15	5.340915e-16	1.683437e-15	-3.660091e-16	-1.227390e-16	88.349619	0.001727
std	47488.145955	1.958696e+00	1.651309e+00	1.516255e+00	1.415869e+00	1.380247e+00	1.332271e+00	1.237094e+00	1.194353e+00	1.098632e+00	...	7.345240e-01	7.257016e-01	6.244603e-01	6.056471e-01	5.212781e-01	4.822270e-01	4.036325e-01	3.300833e-01	250.120109	0.041527
min	0.000000	-5.640751e+01	-7.271573e+01	-4.832559e+01	-5.683171e+00	-1.137433e+02	-2.616051e+01	-4.355724e+01	-7.321672e+01	-1.343407e+01	...	-3.483038e+01	-1.093314e+01	-4.480774e+01	-2.836627e+00	-1.029540e+01	-2.604551e+00	-2.256568e+01	-1.543008e+01	0.000000	0.000000
25%	54201.500000	-9.203734e-01	-5.985499e-01	-8.903648e-01	-8.486401e-01	-6.915971e-01	-7.682956e-01	-5.540759e-01	-2.086297e-01	-6.430976e-01	...	-2.283949e-01	-5.423504e-01	-1.618463e-01	-3.545861e-01	-3.171451e-01	-3.269839e-01	-7.083953e-02	-5.295979e-02	5.600000	0.000000
50%	84692.000000	1.810880e-02	6.548556e-02	1.798463e-01	-1.984653e-02	-5.433583e-02	-2.741871e-01	4.010308e-02	2.235804e-02	-5.142873e-02	...	-2.945017e-02	6.781943e-03	-1.119293e-02	4.097606e-02	1.659350e-02	-5.213911e-02	1.342146e-03	1.124383e-02	22.000000	0.000000
75%	139320.500000	1.315642e+00	8.037239e-01	1.027196e+00	7.433413e-01	6.119264e-01	3.985649e-01	5.704361e-01	3.273459e-01	5.971390e-01	...	1.863772e-01	5.285536e-01	1.476421e-01	4.395266e-01	3.507156e-01	2.409522e-01	9.104512e-02	7.827995e-02	77.165000	0.000000
max	172792.000000	2.454930e+00	2.205773e+01	9.382558e+00	1.687534e+01	3.480167e+01	7.330163e+01	1.205895e+02	2.000721e+01	1.559499e+01	...	2.720284e+01	1.050309e+01	2.252841e+01	4.584549e+00	7.519589e+00	3.517346e+00	3.161220e+01	3.384781e+01	25691.160000	1.000000

	Amount	Amount_log
0	149.62	5.014760
1	2.69	1.305626
2	378.66	5.939276
3	123.50	4.824306
4	69.99	4.262539
5	3.67	1.541159
6	4.99	1.790091
7	40.80	3.732896
8	93.20	4.545420
9	3.68	1.543298

	Time	Time_hours
0	0.0	0.000000
1	0.0	0.000000
2	1.0	0.000278
3	1.0	0.000278
4	2.0	0.000556
5	2.0	0.000556
6	4.0	0.001111
7	7.0	0.001944
8	7.0	0.001944
9	9.0	0.002500

	Time	Time_hours
count	284807.000000	284807.000000
mean	94813.859575	26.337183
std	47488.145955	13.191152
min	0.000000	0.000000
25%	54201.500000	15.055972
50%	84692.000000	23.525556
75%	139320.500000	38.700139
max	172792.000000	47.997778

	Time	V16	Amount	V28	V27	V26	V25	V24	V23	V22	...	V10	V9	V8	V7	V6	V5	V4	V3	V2	Class
Total	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Pourcentage	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

	Skewness	Interprétation
Time	-0.035568	≈ 0 (distribution quasi symétrique)
Amount	16.977724	>> 0 (forte asymétrie positive)

	Nombre	Pourcentage (%)	Interprétation
Class 0	284315	99.83	Classe majoritaire (transactions légitimes)
Class 1	492	0.17	Classe minoritaire (transactions frauduleuses)

	Skewness	Interprétation
Amount	16.977724	>> 0 : forte asymétrie positive
Amount_log	0.162703	≈ 0 : distribution quasi symétrique

	Amount	Amount_log	Amount_scaled	Time_hours	Time_period
0	149.62	5.014760	1.124303	0.000000	Nuit
1	2.69	1.305626	-1.114639	0.000000	Nuit
2	378.66	5.939276	1.682368	0.000278	Nuit
3	123.50	4.824306	1.009339	0.000278	Nuit
4	69.99	4.262539	0.670241	0.000556	Nuit

	Precision (Fraude=1)	Recall (Fraude=1)	F1-score (Fraude=1)	ROC-AUC
Méthode
Tomek Links	0.850	0.649	0.736	0.953
SMOTE	0.094	0.865	0.169	0.966

Détection de la fraude par carte de crédit – Modèles prédictifs

Pipeline complet de préparation des données et d’apprentissage automatique

Interprétation des distributions et de l’asymétrie¶

Interprétation des boxplots¶

Interprétation de la répartition des classes¶

Interprétation de la densité temporelle des transactions¶

Interprétation de la transformation logarithmique¶

Interprétation des résultats – Tomek Links¶

Interprétation de la matrice de confusion – Tomek Links¶

Interprétation des résultats – SMOTE¶

Interprétation de la matrice de confusion – SMOTE¶

Interprétation des résultats¶

🔹 Vérification de la structure du dataset final¶

🔹 Vérification des types de variables¶

🔹 Vérification des valeurs manquantes¶

🔹 Vérification des statistiques descriptives¶

🔹 Vérification de l’asymétrie¶

Interprétation de la distribution de Amount_scaled¶

🔹Matrice de corrélation¶

Interprétation de la matrice de corrélation¶